agentleFS
alexmonegropalexmonegrop/job-hunt-os/.claude/skills/session-start/SKILL.md

session-start

Skill0 starsChanged 7 months agoMITView on GitHub

Start-of-session bootup: select active user, test MCP connectivity (database, LinkedIn), check MEMORY.md for priorities, review last session summary, verify tool health, and present a prioritized action menu. Use at the beginning of every work session.

  • Reads credentials

42 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use deploy-verify, reflect and 2 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • danielarif26Skill

    deploy-verify

    Verify a Vercel deployment actually went live and serves the correct production domain. Self-improving: reads its own run history before checking and logs the outcome after. Use when the user asks "did you deploy?", "is it live?", "verify the deploy", "check vercel", or after any Vercel deploy. Part of the Agentic OS (Level 1 backbone).

    Also in 3 projects that use this

  • Microsoft AzureSkill

    reflect

    Learning capture system that extracts HIGH/MED/LOW confidence patterns from conversations to prevent repeating mistakes. Use after user corrections ("no", "wrong"), praise ("perfect", "exactly"), or when discovering edge cases. Complements .squad/agents/{agent}/history.md and .squad/decisions.md.

    Also in 4 projects that use this

  • ibahgatSkill

    autopilot

    Run autonomous multi-stage execution loops until acceptance criteria pass or blockers are reached.

    Also in 3 projects that use this

  • MicrosoftSkill

    handoff

    Compact the current conversation into a handoff document for another agent to pick up.

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. 0 are common, 10 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 35 edited versions we read keep these.

  • 10

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    ALWAYS start by identifying who is using the system.

    line 25
  2. Distinctive:

    If the user picks an existing person:

    line 40
  3. Very specific:

    Use their user_id for all DB queries this session.

    line 43
  4. Distinctive:

    If the user picks "New user":

    line 45
  5. Distinctive:

    After onboarding completes, continue with normal session start for the new user.

    line 47
  6. Distinctive:

    Read MEMORY.md (already in system prompt for Claude Code, but review the "Next Session Priorities" and status sections explicitly).

    line 53
  7. Very specific:

    Read the overall plan: plans/OVERALL-PLAN.md (if present).

    line 61
  8. Distinctive:

    If no session summary or plan found, note it and continue.

    line 70
  9. Distinctive:

    If the database MCP is connected, run:

    line 89
  10. Distinctive:

    ALWAYS pull latest from GitHub before starting work — the repo can be the shared source of truth across users / sessions.

    line 119
  11. Very specific:

    If using PAT-in-URL pattern (when SSH keys / credential cache aren't available), use the documented "set token → operation → strip token" …

    line 125
  12. Distinctive:

    If succeeds: "Git: synced with remote"

    line 127
  13. Distinctive:

    If conflicts: "Git: CONFLICTS — resolve before proceeding" (show conflicting files)

    line 128
  14. Very specific:

    If 401: "Git: PAT expired — regenerate per docs/SETUP.md"

    line 129
  15. Distinctive:

    If network failure: "Git: pull failed — working offline, will sync later"

    line 130

“Common” is measured: kept by at least 6 in 10 of the 35 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Session Start

Standardized session bootup to quickly orient and verify all systems are working.

What This Skill Does

  1. Select active user — ask who is using the system (or onboard a new user).
  2. Read MEMORY.md — load current project state and priorities.
  3. Find last session summary — check what was done and what's next.
  4. Test MCP connectivity — verify database, LinkedIn, Context7, Playwright.
  5. Quick database health check — confirm base access, list tables.
  6. Present action menu — show prioritized options for this session.
  7. Git pull (latest) — sync with remote.

Workflow

Step 0: Select active user

ALWAYS start by identifying who is using the system. Read JOB_HUNT_USER from .env. If unset or set to a value not present in the database users table, proceed with selection:

mcp__nocodb__list_records({
  base_id: "${NOCODB_BASE_ID}",
  table_name: "users",
  fields: "id,full_name,slug,location,target_roles",
  limit: 10
})

Present an AskUserQuestion with:

  • One option per existing user (show name + location).
  • A "New user (onboard someone new)" option.

If the user picks an existing person:

  • Set that person as the active user for the session.
  • Load their applications/resumes/data/{slug}/user-config.yaml.
  • Use their user_id for all DB queries this session.

If the user picks "New user":

  • Invoke /onboard-user — the full automated onboarding flow.
  • After onboarding completes, continue with normal session start for the new user.

Store the active user slug in session context. All subsequent skills and queries should use this user's data.

Step 1: Load context

Read MEMORY.md (already in system prompt for Claude Code, but review the "Next Session Priorities" and status sections explicitly).

Find and read the most recent session summary for this user:

Glob for: plans/archive/SESSION-SUMMARY-{user-slug}-*.md
Read the most recent one (by date in filename).

Read the overall plan: plans/OVERALL-PLAN.md (if present). This is the living master plan with workstream status, metrics, blockers, and next steps.

Find and read any active sub-plan (if any):

Check MEMORY.md and OVERALL-PLAN.md for "IN PROGRESS" references.
Glob for: plans/**/*.md (excluding archive/) — find plans with incomplete steps.
Read any plan with incomplete steps (⬜ markers or "Resume at" notes).

If no session summary or plan found, note it and continue.

Step 2: Test MCP connectivity (parallel)

// Database (NocoDB by default)
mcp__nocodb__list_tables({ base_id: "${NOCODB_BASE_ID}" })

// LinkedIn
mcp__linkedin__get_company_profile({ company_name: "anthropic" })

// Context7 (if configured)
mcp__context7__resolve_library_id({ libraryName: "python" })

Report results as a status table.

Step 3: Database health check

If the database MCP is connected, run:

mcp__nocodb__list_tables({
  base_id: "${NOCODB_BASE_ID}",
  detailLevel: "tableIdentifiersOnly"
})

Report:

  • Number of tables found
  • Whether job_postings and applications tables exist
  • Any tables with unexpected names

Step 4: Check pending must-dos

Review MEMORY.md "Next Session Must-Dos". For each item: mark DONE / PENDING / BLOCKED.

Step 5: Present action menu

Based on context gathered, present the user with a prioritised action menu via AskUserQuestion:

  1. Infrastructure setup — if tables need creating, MCPs need fixing, etc.
  2. Networking outreach — cold outreach, contact population, follow-ups
  3. Job search — run job searches, review postings, score opportunities
  4. Application prep — tailor resumes, draft cover letters, prepare applications
  5. Meeting prep — prepare for upcoming meetings/interviews
  6. Free-form — user has a specific task in mind

Step 6: Git pull (sync with remote)

ALWAYS pull latest from GitHub before starting work — the repo can be the shared source of truth across users / sessions.

git pull --rebase origin main

If using PAT-in-URL pattern (when SSH keys / credential cache aren't available), use the documented "set token → operation → strip token" pattern from docs/SETUP.md. NEVER leave a PAT in the remote URL.

  • If succeeds: "Git: synced with remote"
  • If conflicts: "Git: CONFLICTS — resolve before proceeding" (show conflicting files)
  • If 401: "Git: PAT expired — regenerate per docs/SETUP.md"
  • If network failure: "Git: pull failed — working offline, will sync later"

Report git sync status in the output table alongside MCP status.

Output Format

## Session Start — YYYY-MM-DD
**Active User**: [Name] (user_id: X, slug: [slug])

### MCP Status
| Service | Status | Notes |
|---------|--------|-------|
| Database | ✅ / ❌ | [details] |
| LinkedIn | ✅ / ❌ | [details] |
| Context7 | ✅ / ❌ | |
| Playwright | ✅ / ❌ | |
| Git | ✅ / ❌ | |

### Database Tables
[List of tables]

### Last Session ([date])
**Done**: [brief]
**Next**: [priorities]

### Overall Plan
**Workstreams**: [summary]
**Blockers**: [active blockers]

### Active Sub-Plan (if any)
**Plan**: [name + path]
**Resume at**: [Phase X, Step Y]
**Remaining**: [count]

### Pending Must-Dos
- [ ] / [x] Item 1
- [ ] / [x] Item 2

### What would you like to work on?
[AskUserQuestion]

Rules

  1. Run MCP tests in parallel — don't test sequentially.
  2. Don't block on MCP failures — report status and continue.
  3. Keep output concise — status table, not paragraphs.
  4. Always end with action menu — user picks what to work on.
  5. If LinkedIn MCP fails, suggest checking Docker — most common issue.
  6. If database MCP fails, suggest checking credentials — second most common.
  7. Git pull BEFORE any work — another user / session may have pushed changes.
  8. Never leave a PAT in the remote URL.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.